Section 230 Explained: The Law Behind the Internet

• What is Section 230?
• How does Section 230 work?
• Who does Section 230 protect?
• Why Section 230 matters (even if you hate social media)
• Section 230 in real cases: what courts have actually done
• Section 230 repeal, Section 230 reform, and Section 230 news
• Section 230 and AI: does CDA 230 cover chatbots?
• What you can do (practical steps)
• FAQ

Section 230 (also called Section 230 of the Communications Decency Act or CDA 230) is a U.S. law that generally says online services and their users aren’t legally treated as the “publisher or speaker” of content posted by someone else. In plain terms: if another person posts something on a site or app, Section 230 often blocks lawsuits that try to blame the site for that third-party content.

What is Section 230?

Section 230 is part of the Communications Decency Act, passed as Title V of the Telecommunications Act of 1996, and codified at 47 U.S.C. § 230. You can read the statute text at Cornell Law School’s 47 U.S.C. § 230 page.

People often call it “the 26 words” because of the core sentence in § 230(c)(1):

“No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.”

It was authored by Rep. Christopher Cox (R-CA) and Rep. Ron Wyden (D-OR). Legal writer Jeff Kosseff popularized the “26 words” framing in his 2019 book title, The Twenty-Six Words That Created the Internet (the idea, not a legal rule).

How does Section 230 work?

At a high level, Section 230 does two related things: it limits liability for third-party content, and it protects “good faith” moderation.

1) The “publisher/speaker” shield (the famous 26 words)

Under § 230(c)(1), a provider (and also a user) of an “interactive computer service” generally can’t be treated as the publisher or speaker of content that came from “another information content provider.” That matters because many lawsuits against platforms are basically “publisher liability” claims—defamation-style theories, negligence theories, and other claims that depend on treating the service like it published the harmful statement.

The foundational early appellate case is Zeran v. America Online (4th Cir. 1997). The court read § 230 broadly, holding it bars “any cause of action that would make service providers liable for information originating with a third-party user.”

2) The “Good Samaritan” moderation protection

§ 230(c)(2) is often described as the “Good Samaritan” provision. It provides protection for removing or restricting content the provider considers objectionable, in good faith. This is part of why sites can take down posts (or filter categories of posts) without automatically becoming liable for what they didn’t take down.

What Section 230 does not do (important limits)

Even though Section 230 is broad, it has clear boundaries in the law as described in the statute and related updates:

• It does not protect content the platform itself creates. If the service is the “information content provider” for a piece of content, § 230’s third-party shield doesn’t apply to that content.
• It does not shield platforms from federal criminal law.
• It does not apply to intellectual property claims.
• Since 2018, it does not provide full protection for certain sex-trafficking claims because of FOSTA-SESTA (2018).

Those limits are a big part of why Section 230 reform debates can get confusing: some harms people care about are already outside § 230, while others are often blocked by it depending on how the claim is framed.

Who does Section 230 protect?

The short answer to “who does Section 230 protect?” is: it protects both platforms and users.

That “and users” part is easy to miss. § 230(c)(1) applies to “provider or user of an interactive computer service.” In practice, that can matter for ordinary people who:

• Run a neighborhood forum, Discord, or small community group
• Host comments on a blog
• Moderate a volunteer-run nonprofit site
• Operate a niche marketplace with listings from other people

This is why many civil-liberties-focused groups argue that a broad Section 230 repeal wouldn’t just hit “Big Tech,” but could expose smaller sites and even individual moderators to expensive legal threats. The Electronic Frontier Foundation makes this user-side argument prominently on its CDA 230 issue page.

Why Section 230 matters (even if you hate social media)

Even if you don’t use the biggest social platforms, Section 230 quietly shapes what the internet looks like day to day—what small communities can exist, how much moderation happens, and who can afford to host public speech.

It was designed to avoid punishing moderation

One of the key historical motivations was to override Stratton Oakmont v. Prodigy (N.Y. 1995). That case held that a platform that moderated content could be treated as a publisher and sued for it—creating an incentive to do no moderation to avoid liability. Section 230 was written to stop that “moderate and get punished” dynamic.

Repeal would change the risk for anyone who hosts other people’s words

The practical concern flagged in the reform debate is straightforward: if Section 230 repeal exposed hosts to liability for user posts, many sites would respond by (1) pre-screening posts, (2) using heavier automated filtering, (3) taking down more lawful speech to reduce risk, or (4) shutting down comments and communities entirely—especially smaller sites that can’t absorb litigation risk.

Whether you view that as a “feature” or a “bug” depends on what harms you prioritize. But it’s not an abstract legal fight: it can translate into which communities survive online.

It intersects with today’s online-safety push

In the U.S., online-safety proposals often connect—directly or indirectly—to platform liability. If you’re tracking debates about age checks and identity systems, it helps to read Section 230 in the same mental map as:

• Ban the Bots: age verification laws
• Ban the Bots: digital ID
• Ban the Bots: AI regulation

Section 230 in real cases: what courts have actually done

To understand what Section 230 does in real life, it helps to look at what courts and lawmakers have actually done with it—especially in the last few years.

Zeran v. AOL (1997): the broad early interpretation

Zeran v. America Online (4th Cir. 1997) is widely recognized as the first major appellate decision interpreting § 230. The ruling established a broad immunity approach, barring claims that would make services liable for third-party information.

This case is one reason Section 230 news often focuses on victims’ inability to sue platforms: Zeran is the “big early door” that courts walked through to apply § 230 broadly.

Gonzalez v. Google (2023) and Twitter v. Taamneh (2023): the Supreme Court sidestep

In Gonzalez v. Google (2023), the U.S. Supreme Court declined to decide the key Section 230 question and sent the case back after deciding a companion case.

That companion was Twitter v. Taamneh (2023), where the Court unanimously held platforms did not “aid and abet” terrorism under the legal theory at issue. That ruling made the Gonzalez § 230 question unnecessary for the Court to resolve then.

The takeaway: even in major cases, the Supreme Court has not provided a clean, modern “this is exactly how § 230 applies to today’s platforms” reset.

Moody v. NetChoice (July 1, 2024): moderation is protected speech

Moody v. NetChoice (July 1, 2024) isn’t a Section 230 case in the narrow sense, but it’s central to the same fight: platform responsibility and moderation. The Court (6–3, in an opinion by Justice Kagan) held that platforms’ content-moderation choices are protected by the First Amendment and that there is “no social media exception to the First Amendment.”

This matters for Section 230 reform because some proposals try to pressure platforms into carrying certain speech or moderating differently. Moody is a reminder that even if Congress changes § 230, the First Amendment still shapes what governments can force platforms to do.

Garcia v. Character Technologies (M.D. Fla., May 21, 2025): a key Section 230 and AI warning shot

Garcia v. Character Technologies (Middle District of Florida, May 21, 2025) is a wrongful-death suit arising after the suicide of 14-year-old Sewell Setzer III. Judge Anne Conway largely denied the motion to dismiss, treated Character.AI as a “product,” and declined to hold that chatbot output is fully protected speech.

This case is one reason “section 230 and ai” has become a common search: generative systems don’t fit neatly into a law written for hosting other people’s posts.

For more on litigation pushing these boundaries, see Ban the Bots: AI lawsuits.

Section 230 repeal, Section 230 reform, and Section 230 news

Calls for Section 230 repeal and Section 230 reform have become politically bipartisan—but for different reasons.

Why both parties criticize CDA 230 (but not for the same reason)

• Many Democrats argue Section 230 lets platforms avoid accountability for harms like hate speech and disinformation.
• Many Republicans argue it shields platforms that “censor” conservative viewpoints.

Those motives can point toward very different policy solutions—even when the slogan sounds the same.

Major recent legislative effort: the Sunset Section 230 Act

In December 2024, Sens. Dick Durbin (D-IL) and Lindsey Graham (R-SC) introduced the “Sunset Section 230 Act,” designed to repeal § 230 two years after enactment as leverage to force renegotiation.

On December 18, 2025, it was reintroduced as S.3546 (119th Congress) by Judiciary Chairman Chuck Grassley with Graham and Durbin. The cosponsors span both parties: Whitehouse, Hawley, Klobuchar, Blackburn, Blumenthal, Moody, Welch. A House companion, H.R.6746, was also introduced. You can track the Senate bill here: S.3546 on Congress.gov.

Presidents and Section 230

Section 230 has also been a target at the executive-branch messaging level:

• Donald Trump issued a 2020 executive order titled “Preventing Online Censorship” seeking to narrow § 230 and has repeatedly called for repeal.
• Joe Biden said in 2020 that § 230 “should be revoked, immediately.”

A quick comparison: repeal vs targeted reform

Not everyone who wants change wants the same kind of change. Here’s a plain-English comparison of the three main camps described in the public debate (as summarized in the research context, including views associated with EFF and reform-oriented groups like Public Knowledge and Brookings).

Comparison: Section 230 repeal vs Section 230 reform

• Full repeal / sunset: Removes the liability shield broadly. Supporters argue victims should be able to sue platforms over harms like harassment, illegal drug sales, CSAM-related harms, sextortion, and defamation. Critics argue it would push over-removal and favor the biggest firms that can afford lawsuits.
• Narrow, targeted reforms: Keeps the core but adjusts specific carveouts or standards. Middle-ground reformers (including positions associated with Public Knowledge and Brookings, per the research context) generally prefer this approach over full repeal.
• Keep it largely as-is: The EFF-style argument emphasizes that § 230 protects users and small sites—not just Big Tech—and that weakening it could cement Big Tech dominance and encourage upload filters and sweeping takedowns of lawful speech.

Section 230 and AI: does CDA 230 cover chatbots?

Section 230 and AI is where the law’s original language runs into a modern reality: generative AI systems can produce new text, images, and audio that look “authored,” not merely hosted.

The key unresolved question: who is the “information content provider”?

Section 230’s core protection applies when the information was “provided by another information content provider.” With many generative AI tools, the argument is that the system (or its developer/operator) may be the one providing the information—meaning the output may fall outside § 230.

This is why many commentators argue Section 230 “almost certainly” does not cover AI-generated content—but, as the research context notes, there is no controlling appellate ruling that conclusively settles it across the U.S.

What legal experts are saying (in plain English)

Fordham law professor Chinmayi Sharma sums up the core intuition many people have: “Section 230 was built to protect platforms from liability for what users say, not for what the platforms themselves generate.” In that framing, chatbot output “looks far less like neutral intermediation and far more like authored speech.”

That distinction matters for real harms tied to synthetic content—like scams, defamation, and deepfakes. If you’re following how AI-created content spreads, see:

• Ban the Bots: deepfakes
• Ban the Bots: Grok deepfakes

Legislation aimed directly at AI and Section 230

In 2023, Sen. Josh Hawley introduced the “No Section 230 Immunity for AI Act,” which would exclude generative AI expressly. (The point here isn’t whether it will pass; it’s that lawmakers are explicitly targeting the “Section 230 and AI” gap.)

AI liability cases that may test Section 230

Two cases highlighted in the research context show how this could play out:

• Garcia v. Character Technologies (M.D. Fla., May 21, 2025): The court largely denied a motion to dismiss and treated Character.AI as a “product,” declining to hold chatbot output is fully protected speech.
• Raine v. OpenAI (filed August 2025, San Francisco): A lawsuit over the suicide of 16-year-old Adam Raine that names OpenAI and CEO Sam Altman. The research context notes it may become an early test of Section 230’s reach over AI.

A reputable recent summary of this uncertainty is Fortune’s October 2025 reporting on AI chatbots and Section 230.

What you can do (practical steps)

If you’re not a lawyer or a policy professional, you can still respond to Section 230 news in concrete ways—without having to pick a “team” for or against Big Tech.

1. Track the actual bills (by number), not just the headlines. The current “sunset” proposal is S.3546 (with House companion H.R.6746). Read summaries and status updates on Congress.gov’s S.3546 page.
2. Ask your representative a specific question: “If Section 230 is weakened or repealed, what protections will exist for small forums, nonprofits, and individual moderators who host third-party content?” This gets at the “who does Section 230 protect” issue, not just Big Tech talking points.
3. Read a civil-liberties view of CDA 230. The EFF’s CDA 230 issue page is a clear starting point for understanding the user-side stakes.
4. If AI harms are what brought you here, follow the lawsuits. Courts are actively grappling with whether chatbot outputs are more like “hosting” or more like “creating.” Ban the Bots tracks this area at /ai-lawsuits/.
5. Connect Section 230 to the broader AI and platform-policy picture. If you care about how the internet is changing—jobs, enforcement, infrastructure—see /explainers/ai-regulation and our ongoing coverage of real-world impacts at /ai-backlash/.

FAQ

Is Section 230 a “Big Tech loophole”?

Section 230 certainly benefits large platforms, but the statute text also protects users and any service hosting third-party content. The EFF and other defenders argue that removing it could hit small sites hardest, because they can’t afford legal exposure and would be pushed toward heavy filtering or closure.

Does Section 230 mean platforms can never be sued?

No. Section 230 does not shield platforms from federal criminal law or intellectual property claims, and since FOSTA-SESTA (2018) it does not provide full protection for certain sex-trafficking-related claims. It also doesn’t protect content the platform itself creates.

Did the Supreme Court “end Section 230” in 2023?

No. In Gonzalez v. Google (2023), the Supreme Court declined to decide the Section 230 question. The companion decision in Twitter v. Taamneh (2023) made it unnecessary for the Court to resolve the 230 issue in that moment.

What’s the difference between Section 230 and the First Amendment?

Section 230 is a statute about liability for third-party content and moderation. The First Amendment is a constitutional rule limiting government restrictions on speech. In Moody v. NetChoice (July 1, 2024), the Court emphasized that platforms’ moderation choices are protected expression and there is “no social media exception to the First Amendment.”

Does Section 230 cover AI-generated chatbot answers?

That’s disputed and unresolved at a controlling appellate level. The key question is whether the output is “information provided by another information content provider,” or whether the AI system/operator is effectively the content provider. In Garcia v. Character Technologies (M.D. Fla., May 21, 2025), the court largely denied a motion to dismiss and treated the chatbot as a “product,” signaling that chatbot outputs may not be treated as fully protected speech.

What should I watch in Section 230 reform right now?

Watch concrete legislation like the “Sunset Section 230 Act” reintroduced as S.3546 (with House companion H.R.6746) and track it on Congress.gov. Also watch how AI-related cases develop—like Raine v. OpenAI (filed August 2025)—because courts’ reasoning about “who created the content” could reshape how Section 230 works in practice.

Conclusion: why Section 230 is still the law behind the internet

Section 230—Section 230 of the Communications Decency Act, or CDA 230—is still the backbone rule that determines when online services (and users) are liable for content posted by someone else. That’s why Section 230 news, Section 230 reform, and Section 230 repeal debates matter for ordinary people: they can change whether small communities can host speech at all, and how accountability works when harm happens—especially as section 230 and ai questions move into the courts.

If you want to keep learning and take action, start here: follow the real-world impact threads at /ai-backlash/, explore cases at /ai-lawsuits/, and if you’re worried about where this is all heading for work and communities, see /ai-layoffs/, ways to respond at /fighting-back/, and the infrastructure driving the shift at /data-center-map/.